A Modified Neural Network for Improving River Flow Prediction/Un Réseau de Neurones Modifié pour Améliorer la Prévision de L'Écoulement Fluvial

Abstract

Artificial neural network (ANN) models provide huge potential for simulating nonlinear behaviour of hydrological systems. However, the potential of ANN is yet to be fully exploited due to the problems associated with improving the model generalization performance. Generalization refers to the ability of a neural network to correctly process input data that have not been used for calibrating the neural network model. In the hydrological context, better generalization performance implies higher precision of forecasting. The primary objectives of this study are to explore new measures for improving the generalization performance of an ANN-based rainfall–runoff model, and to evaluate the applicability of the new measures. A modified neural network model (entitled goal programming (GP) neural network) for modelling the rainfall–runoff process has been developed, in which three enhancements are made as compared to the widely-used backpropagation (BP) network. The three enhancements are (a) explicit integration of hydrological prior knowledge into the neural network learning; (b) incorporation of a modified training objective function; and (c) reduction of network sensitivity to input errors. Seven watersheds across a range of climatic conditions and watershed areas in China were selected for examining the alternative networks. The results demonstrate that the GP consistently outperformed the BP both in the calibration and verification periods and three proposed measures yielded improvement of performance.     

Application of chloride profile and water balance methods in estimating groundwater recharge in Luanjing Irrigation Area,Inner Mongolia / Application des méthodes du profil de chlorure et du bilan hydrique pour l'estimation de la recharge hydrogéologique dans le Périmètre Irrigué de Luanjing,Mongolie Intérieure

Abstract

A study was carried out to investigate the use of the chloride profile method in conjunction with the water balance method to estimate the annual groundwater recharge in both natural and irrigation sites in Luanjing Irrigation Area, Inner Mongolia. Groundwater recharge from precipitation, estimated by the chloride profile method, is less than 0.1 mm year^?1 which accounts for just 0.06% of the long-term average annual rainfall, indicating that rainfall presently plays a minor role in the groundwater recharge. It appears that recharge events only occurred after heavy rain or sustained rainfall events. In the cropped area, the chloride profile method indicated that the average annual recharge is 268 mm year^?1 with an infiltration rate of 32.5%, which is reasonably consistent with the 33.1% obtained by the water balance method in 2007. The study shows that about one third of that water is discharged back to the groundwater.     

Influences of coupled wind–water processes on suspended sediment grain size: an example from tributaries of the Yellow River / Influences de processus couplés éoliens-hydriques sur la granulométrie des matières en suspension: exemple d'affluents du Fleuve Jaune

Abstract

Some unique coupled wind–water erosion processes exist in the desert-loess transitional zone in the middle Yellow River basin. Based on data from 40 stations on 29 rivers, a study was made on the influence of such processes on suspended sediment grain-size characteristics of the tributaries of the Yellow River. Results show that the percentage of >0.05-mm grain size decreases with the increased annual mean precipitation, but increases with the increase in the annual mean number of sand-dust storm days. The percentage of <0.01-mm grain size increases with the increase in the annual mean precipitation, but decreases with the increase in the annual number of sand-dust storm days. Based on annual mean data from 40 stations, multiple regression equations were established between the percentages of >0.05-mm grain size (r _>0.05) and <0.01-mm grain size (r _<0.01), annual mean precipitation (P _m) and annual mean number of sand-dust storm days (D _ss). On this basis, the relative contributions of the variations in D _ss and P _m to the variation in r _>0.05 and r _<0.01 were estimated. The results indicate that the variation in sand-dust storm frequency exerts greater influences on the variation in grain-size characteristics of suspended load than does the variation in annual mean precipitation. With the increase in the coupled wind–water processes index, expressed by P _m/D _ss, the percentage of >0.05-mm grain size in suspended sediment decreases and the percentage of <0.01-mm grain size increases. With the variation in P _m/D _ss, different combinations of r _>0.05 with r _<0.01 appear, which have some influence on the formation of hyperconcentrated flows. There exist some optimal ratios of coarse to fine fractions in suspended sediment that make sediment concentrations of hyperconcentrated flows the highest. The optimal r _>0.05/r _<0.01 value is related to some range of the index P _m/D _ss. When the P _m/D _ss index falls in this range, the optimum combination of relative coarse with fine sediments in the suspended load appears, and thus results in the peak values of sediment concentration.